Use of magnetic resonance angiography for the preoperative evaluation of patients with infrainguinal arterial occlusive disease

Use of magnetic resonance angiography for the preoperative evaluation of patients with infrainguinal arterial occlusive disease John R. Hoch , MD, Michael J. Tullis, MD, T o d d W. Kennell , MD, John McDermot t , MD, Charles W. Acher, MD, and William D. Turnipseed, MD, Madison, Wis.

Purpose: This study was designed to determine whether magnetic resonance angiography (MRA) will allow preoperative management decisions without the need for contrast arteriography in patients with lower extremity ischemia caused by infrainguinal arterial occlusive disease. Methods: Forty-five patients with lower extremity ischemia in 50 limbs were evaluated by both two-dimensional time-of-flight MRA and intraarterial digital subtraction angiography (DSA) between February 1992 and June 1995. Independent management plans were based on clinical presentation, pulse volume recordings, and separate reviews of the MRA and DSA. Results: Of 50 limbs, 23 required arterial bypass, 19 percutaneous transluminal angioplasty, 5 patch angioplasty, and 3 amputation. MRA and DSA correlated exactly in 89.5% of infrainguinal arterial segments, whereas interpretations disagreed in 10.5% of arterial segments. Mismatches that had an influence on patient treatment decisions occurred in only 8 (2.3%) of 352 arterial segments. Independent MRA- and DSA-based revascularization plans agreed in 45 (90%) extremities. MRA predicted the level of arterial reconstruction in all 23 limbs that required arterial bypass. MRA identified focal stenoses amenable to percutaneous transluminal angioplasty in 18 (94.7%) of the 19 limbs that ultimately underwent percutaneous transluminal angioplasty. A strategy of preoperative planning by MRA with confirmatory intraoperative arteriography would represent a 31% cost savings per patient at our institution while eliminating the morbidity of preoperative DSA. Conclusions: When used in combination with the patient's physical examination and segmental limb pressures with plethysmography, MRA is sufficient for planning infrainguinal arterial bypass procedures and selecting patients for percutaneous transluminal angioplasty. (J Vasc Surg 1996;23:792-801.)

Successful infrainguinal bypass surgery is dependent on the adequate angiographic visualization of the lower extremity arteries to plan proximal and distal sites of anastomosis and to assess the extent of the occlusive process. The preoperative gold standard at our institution has been conventional angiography in combination with digital subtraction angiography

From the Section of Vascular Surgery, Department of Surgery, University of Wisconsin.

Presented at the Nineteenth Annual Meeting of the Midwestern Vascular Surgical Societ); Chicago, Ili., Sept. 22-23, 1995.

Reprint requests: John R. Hoch, MD, Department of Surgery- H4/736, University of Wisconsin, 600 Highland Ave., Madi- son, WI 53792.

Copyright © 1996 by The Society for Vascular Surgery and International Society for Cardiovascular Surgery, North Ameri- can Chapter.

0 7 4 1 - 5 2 1 4 / 9 6 / $ 5 . 0 0 + 0 2 4 / 6 / 7 2 4 2 6

792

(DSA)) Data pooled from several studies reveal that distal arteries are adequately visualized with standard arteriography in only 75% to 83% of patients. 2 Several techniques have bccn advocated to improve visualization of distal arteries including DSA, delayed filming, reactive hypcremia, vasodilator injection, external warming, and selective catheterization. With these special techniques visualization o f distal arteries im- proves to between 71% and 100% of extremities. 2 Conventional angiography, however, carries with it the contrast-related risks of renal failure and allergic reactions and the risk of complications caused by the arterial puncture. 2-4

Magnetic resonance angiography (MRA) is a noninvasive vascular imaging technique that over the last 10 years has been increasingly used for the

JOURNAL OF VASCULAR SURGERY Volume 23, Number 5 Hoch etal. 793

evaluation of lower extremity ischcmia. S-H Experi- ence with MRA of peripheral runoff arteries at other institutions has suggested that MRA is as sensitive as contrast angiography for the diagnosis of lower extremity arterial occlusive disease and may be more sensitive than conventional angiography for the detection of patent distal runoff vessels. 6-~° Because of observer variability in the interpretation of peripheral MRA studies, it is necessary for every institution to compare MRA results with their own contrast arteriographic standards before relying on MRA as an independent imaging modality. Our preliminary experience with MRA reported a 75% correlation with DSA for the detection of lower extremity bypass graft hemodynamic failures. ~ This report details our subsequent experience with lower extremity MRA compared with DSA to formulate preoperative treatment decisions for patients with infrainguinal arterial occlusive disease. MR3. combined with clinical parameters can predict in most cases whether patients require operative intervention or are candidates for percutaneous transluminal angioplasty (PTA) without the need for contrast arteriography.

M AT E RI A L AND M E T H O D S

Forty-five patients with lower extremity ischemia in 50 limbs were evaluated by both two-dimensional time-of-flight (2-D TOF) MRA and intraarterial DSA between February 1992 and June 1995. A total of11 women and 34 men with an average age of 65 years were evaluated. Eight patients reported severe claudication in 10 extremities, whereas 37 patients had limb-threatening ischemia categorized into 12 extremities with ischemic rest pain, and 28 had nonhcal- ing ulcers or gangrene. Significant comorbid conditions included diabetes melitis in 64% and renal insufficiency (creatinine level >1.5 mg/d l ) in 38% of patients.

All patients underwent careful physical examination and segmental limb pressures with plethysmography (Parks Flow Lab, Parks Medical Elcctronics, Aloha, Ore.). Patients were selected for inclusion in this study if they had a normal proximal thigh pres- sure, a good palpable femoral pulse, and evidence of significant distal limb ischcmia. The mean anlde/braehial index for claudicators was 0.55 and for paticnts with threatened limbs was 0.44.

Axial 2-D TOF MRA sequences with traveling inferior saturation pulses wcrc performed on a 1.5T GE Signa MRI (GE Medical Systems, Milwaukec, Wis.) or a 1.0T Siemens Impact MRI (Siemens Medical Systems, Iselin, N. J.). The body coil was used from the aortic bifurcation to the knee on all

patients. Transmit-receive head (n = 23) or extremity (n = 17) coils were used for obtaining images of the lcnee to the foot in most extremities, whereas early in our experience images of this portion of the extremity were obtained with the body coil alone (n = 10).

On the GE 1.5T system the repetition time varied from 30 to 45 msec, and slice thickness was 2.9 mm (no interslice gap) in the body coil (32 cm field of view) and 1.8 mm (no interslice gap) in the extremity coil (17 cm field of view). A flip angle of 60 degrees, echocardiography time of 6.9 msec, and a 256 × 128 matrix were used on all scans. The average examination time was 90 minutes with use of both the body and extremity coils.

On the Siemens 1.0T system the repetition time varied from 30 to 45 msec, and slice thickness was 4 mm (with a 1 mm overlap) in the body coil (35 cm field of view) and 3 mm (with a 1 mm overlap) in the head coil (22 cm field of view). A flip angle of 40 degrees, echocardiography time of 10 msec, and a 256 x i28 matrix were used on all scans. The average examination time was 60 minutes with use of the body and head coils. Examination time was faster with the Siemens system, because the head coil obtained simultaneous images of both extremities from the knee down, whereas images of each extremity must be obtained separately with the GE system's extremity coil. The MRA imaging parameters are the recom- mended 2-D TOF protocols for each MRI unit.

Projection angiography images were created from the axial images by using the maximum intensity projection algorithm on the respective MRI units. MICA was the initial study in all patients, followed by intraarterial DSA within 10 days.

intraarterial DSA was performed in all patients with our standardized protocol. 14 Thirty-seven patients underwent DSA aortography with bilateral runoff via a retrograde femoral catheterization, and eight patients had antegrade femoral catheterization so that PTA could be performed on segmental lesions demonstrated by previous MRA studies. Nonionic contrast was used in all cases (range 100 to 200 ml). Confirmatory intraoperative arteriography was performed in all patients who underwent operative revascularization.

All MRA and DSA studies were reviewed in a blinded fashion by two experienced radiologists. The lower extremity arteries of the ischemic limb were divided into eight segments: common femoral, profunda femoris, superficial femoral, popliteal, anterior tibial, posterior tibial, peroneal, and dorsalis pedis arteries. The proximal and distal segments of five grafts were also scored. Vessels were scored as normal,

JOURNAL OF VASCULAR SURGERY 794 Hoch et al. May 1996

stenotic, or occluded. "Normal" arteries had less than 50% luminal narrowings and were considered suitable as recipient or inflow arteries for bypass. "Stenotic" arteries were patent with either focal or multiple hemodynamically significant stenoses (>50% luminal narrowing). "Occluded" referred to completely occluded arteries or arteries that were virtually occluded and not suitable for bypass or angioplasty.

After scoring was performed by the radiologists, MRA and DSA studies were independently reviewed by our group of vascular surgeons. The surgeons scored the MRA studies as to image quality and suitability for preoperative decision malting, that is, whether they believed the MRA alone was sufficient or whether confirmatory DSAwas necessary. After the MRA and DSA studies were independently reviewed, two revascularization plans (MRA vs DSA) were formulated on the basis of each patient's clinical presentation, segmental arterial pressures, and segmental plethysmography, and separate review of the imaging studies. When revascularization plans differed, review of both studies simultaneously led to the performed procedure (Table I).

Comparison was made of the hospital-based costs and professional fees for DSA aortograms with bilateral runoffand the combination of pelvic and bilateral lower extremity 2-D TOF MRA examinations.

RESULTS

Fifty extremities were evaluated by both 2-D TOF MRA and digital subtraction angiography. No MRA or DSAwas judged to be technically inadequate by the radiologists. No contrast-related complications occurred, but one patient required operative repair of the common femoral artery after PTA because of hemorrhage.

Interpretation of MRA and DSA studies correlated exactly in 315 (89.5%) of 352 arterial segments (Table II). Thirty-seven (10.5%) mismatched segments were identified out of the 352 segments viewed by both studies (Tables II and III). The MRA and DSA interpretations disagreed in 28 (13.8%) of 203 infrageniculate arteries compared with only 8 (5.6%) mismatches of 139 suprageniculate arterial segments. No significant differences were seen in the incidence of mismatches between MICA units or coils.

Table III compares the MRA and DSA interpretations. The major disagreement between the studies occurred in the interpretation of "stenotic" segments. Eleven arterial segments that were scored as "stenotic" by MRAwere believed to be less than 50% narrowed on DSA, and conversely, 11 "stenotic" arteries on DSA were found to be "normal" by MRA.

However, an additional 11 arterial segments believed to be occluded on preoperative DSA were found to be patent by MRA ( 1 "normal" and 10 "stenotic") and on subsequent intraopcrativc arteriography. When "normal" and "stenotic" arteries were pooled to- gether as "patent" arteries, the correlation between MRA and DSA improved to 95.7%. Mismatches that had an influence on patient treatment decisions occurred in only 8 (2.3%) of 352 arterial segments, representing 21.6% of the 37 total mismatched segments (Tables II and IV).

Two independent revascularization plans were made for each extremity on the basis of the clinical situation and the separate MRA and DSA images. There was agreement between plans in 45 (90%) extremities. Disagreement between plans occurred in the remaining five (10%) extremities (Table IV). The ultimate procedure performed was based on the MRA in two extremities and on the DSA in the remaining three (Table IV). The surgeons found all the MRA studies to be of good quality and felt comfortable maldng revascularization management decisions on the basis of the combination of MRA and clinical parameters without preoperative DSA in 47 (94%) of 50 extremities. In three cases the MRA and the complexity of the clinical situation led the reviewers to recommend the addition of DSA. In two of these instances (Table IV; patients i and 2) the DSA proved to bc the accurate imaging modality, whereas the MtCA and DSA images correlated in the third.

MRA predicted the level ofartcrial reconstruction in all 23 limbs that required arterial bypass. In one patient MtCA identified a patent tibial artery suitable for bypass that was not seen on DSA (Fig. 1, Table IV; patient 4). In a second patient MRA correctly identified the posterior tibial artery as the site of distal anastomosis, whereas the peroneal artery was interpreted as patent and the posterior tibial occluded by DSA (Fig. 2, Table IV; patient 5). In a third patient with severe cardiac and respiratory compromise, the MRA demonstrated a long anterior tibial artery occlusion (Table IV; patient 2). The plan based on the MRA was to bypass this segment for limb salvage. Subsequent DSA revealed serial stenoses that were treated by PTA in this medically compromised patient. A fourth patient, who was a diabetic with noncompressiblc arteries and poor pulse volume recording waveforms at the digital and metatarsal levels, was found to have an occluded perimalleolar posterior tibial artery on MRA and a severely diseased artery on DSA. Bypass was the planned treatment

JOURNAL OF VASCULAR SURGERY Volume 28, Number 5 Hogh etal. 795

Table I. Proccdures performed and operative indications

Procedure No.

Indications

Claudication Rest pain/tissue loss

Bypass graft Percutaneous transluminal angioplasty Patch angioplasty/endar terectomy Amputation without revascularization Total

23 5 18 19 4 15 5 1 4 3 0 3

50 10 40

Table II. Agreement of MRA with DSA for the detection of infrainguinal segmental arterial occlusive disease

Clinically Exact MRA/DSA MRA/DSA relevant~

Arteries No. of segments match (%) Mismatch (%) mismatches (%)

Common femoral 44 43 (97.7) 1 (2.3) 1 (2.3) Profunda femoris 45 41 (91.1) 4 (8.9) 1 (2.2) Superficial femoral 50 47 (94.0) 3 (6.0) 1 (2.0) Popliteal 50 43 (86.0) 7 (14.0) 0 (0) Anterior tibial 45 38 (84.4) 7 (15.6) 2 (4.4) Posterior tibial 45 39 (86.7) 6 (13.3) 1 (2.2) Peroneal 45 38 (84.4) 7 (15.6) 1 (2.2) Dorsalispedis 18 17 (94.4) 1 (5.6) 0 (0) Grafts* 10 9 (90.0) 1 (10.0) 1 (10) Total 352 315 (89.5) 37 (10.5) 8 (2.3)

*Proximal and distal segments of five bypass grafts. ]""Clinically relevant" mismatch altered preoperative management

based on both imaging modalities; however, a digital amputation was initially performed, and it healed without bypass.

Five extremities underwent cndartercctomy or patch angioplasty of diffusely stenotic arteries or bypass grafts. MRA predicted the appropriate procedure in four of five extremities. In one extremity the MRA failed to predict the appropriate management plan, because surgical clips produced a "signal dropout" near the proximal anastomosis of a bypass graft (Fig. 3). This area was interpreted as an artifact and not a stenosis, because the radiologists knew the clips were present from previous x-ray films (Table II; patient l) . However, this was one of the three cases the surgeons would have performed a preoperative DSA study before operation.

MRA correctly identified focal stenoses amend- able to PTA in 18 (94.7%) of the 19 extremities that underwent PTA. MRA predicted bypass for one extremi~ that was instead managed by PTA (Table II; patient 2). No instances of long-segment occlusions of the superficial femoral artery were seen on DSA that were misread by MRA. One false-positive was the result of an MRA "overread" of a superficial femoral artery stenosis (Table IV; patient 3).

Three diabetic patients underwent amputation without revascularization. All three had noncom-

plan.

Table III . Comparison of normal, stenotic (>50% narrowing), and occluded vessels in 352 segments evaluated separately by MRA and DSA

MRA Evaluation

DSA Evaluation Normal Stenotic Occluded

Normal 155 11 2* Stenotic 11 59 2]" Occluded 15 10 101

*One caused by "in-plane" flow artifact; second was confirmed occluded at surgery. ].Two tibial arteries with serial stenoses. :~Artery occluded by DSA, patent at surgery.

pressible arteries and poor pulse volume recording waveforms in the foot and underwent noninvasive MRA evaluation. MRA revealed no significant stenoses in one patient who had a patent superficial femoral, popliteal, and a peroneal artery open to the ankle and severe arterial occlusive disease of the foot. A second patient was thought to have a superficial femoral stenosis that was not demonstrated on DSA (Table IV; patient 3). The third was the patient described previously with perimalleolar posterior tibial disease who underwent digital amputation before the planned bypass.

Comparison was made of the average costs for

JOURNAL OF VASCULAR SURGERY 796 Hoch et al. May I996

Table IV. Disagreement in five extremities between MRA and DSA-based patient treatment plans

-preoperative plan

-patient DSA MRA Mismatched segments -procedure performed

Proximal and distal anastomotic patch an- gioplasties

2 PTA anterior tibial serial stenoses

3 No PTA/bypass before digital amputation

4 Blind exploration

Popliteal to peroneal bypass; SFA patch an- gioplast3~

Distal anastomotic patch angioplasty

Popliteal to anterior tibial bypass

SFA PTA before digital amputation

DSA: CFA, PFA, and proximal vein graft stenosis MRA: Presumed clip artifact of CFA, PFA, and vein graft DSA: Serial stenoses mid-anterior tibial artery MRA: Mid-anterior tibial 3 cm occlusion DSA: No SFA stenosis MRA: Focal SFA stenosis

Proximal and distal femoral-popliteal anastomotic patch angio- plastics

PTA anterior tibial serial stenoses

No PTA/bypass; digital amputation

Fcmoro-anterior tibial DSA: No reconstitu- Femoro-anterior tibial bypass t_ion of tibial arteries bypass

MRA: Reconstitution of distal anterior tibial artery

Popliteal to posterior DSA: Erroneously cat- tibial bypass; SFA patch egorized peroneal ar- angioplasty tery as patent

MRA: Correctly identified posterior tibial patent with occluded peroneal

Popliteal to posterior final bypass; SFA patch angioplasty

CFA, Common femoral artery; PFA, profunda femoris artery; DSA, digital substraction angiography; MRA, magnetic resonance angiography; .PTA, percutaneous transluminal angioplasw

DSA aortograms with bilateral runoff and the combination of MRA of the pelvis and bilateral lower extremities. Total average costs were $2405 for DSA and $1649 for MRA. Radiology professional fees for DSA were $822 compared with $549 for the MRA studies. Hospital-based costs were $1583 for an outpatient DSA compared with $1100 for an MRA. MRA allowed for a 31% ($756) cost savings per study when MRA of the pelvic arteries with bilateral lower extremity runoff was substituted for the comparable DSA.

DISCUSSION

The purpose of this study was to determine whether MRA can be used as effectively as DSA for treating patients with infrainguinal arterial occlusive disease. Our initial experience with peripheral MRA involved patients with failing infrainguinal bypass grafts, s An exact correlation between MRA and conventional angiography occurred in 75% of cases in this series. Encouraged by this experience, we reasoned that MRA might bc an attractive noninvasivc alterna- tive to conventional angiography for screening patients with lower extremity ischemia. However, for MRA to replace conventional angiography, it would

have to demonstrate a diagnostic accuracy comparable to DSA, allowing surgeons to formulate sound patient treatment plans while being cost-effective and safe.

In this study interpretation of MRA and DSA studies correlated exactly in 89.5% of arterial segments in patients with infrainguinal arterial occlusive disease. Eleven arteries that were believed to be occluded on DSA were found to be "patent" (stenotic or normal) on MRA. Recent experiences with MRA of infrainguinal arteries at other institutions have also demonstrated that MRA is at least as sensitive as contrast angiography for the diagnosis of lower extremity arterial occlusive disease and may be more sensitive than conventional angiography for the detection of patent distal runoff vessels. 610 Carpenter et al.r compared MRA and conventional angiography ofpopliteal and distal runoff arteries in 55 extremities and found that MRA allowed visualization of all the vessels seen on conventional angiography and helped to identify 49% more infrapopliteal arterial segments compared with conventional arteriography. McCau- ley et al. 1° reported identifying 22% more patent arterial segments when MRA images were interpreted by the most experienced reviewer in their group

JOURNAL OF VASCULAR SURGERY Volume 23, Number 5 Hoch et al. 797

Fig. 1. DSA of lower leg failed to identify artery suitable for bypass (A) in this diabetic patient with threatened limb. MRA demonstrated patent distal anterior tibial and dorsalis pedis arteries (B). Intraoperative angiography confirmed preoperative MRA findings (C).

compared with 13% by the least experienced reviewer. Our data suggest that the more severe the proximal occlusive disease is, the more likely it is that MRA will help detect patent tibial arteries not seen on contrast arteriograms. The greater sensitivity of MRA is thought to result from the technique's ability to directly detect flow as slow as 2 cm/sec, unlike conventional arteriography and DSA, which require distal reconstitution of dilute contrast to detect patent arteries.

The surgeons in this study were given the clinical history, noninvasive laboratory findings, and independent MRA and DSA arterial studies from which two patient treatment plans were formulated for each ischemic extremity. MRA and DSA-bascd patient treatment plans correlated in 90% of extremities. In two patients MICA findings improved patient outcomes by identifying patent arteries not sccn on DSA. Carpenter et al.7 reported that MRA revealed additional findings in 48% of limbs, which resulted in altering the intervention plan in 22% of their cases. In 18% of cases MRA identified a runoffvessel for bypass grafting not visualized by conventional arteriography. In a second study Carpenter et al. 6 used outpatient

MRA to plan operative interventions without preoperative arteriography in 67 patients who underwent infrainguinal bypass. 6 Postbypass intraoperative arteriography was used to obtain images of the runoff vessels and confirmed the accuracy of MRA in planning distal bypass surgical procedures in every case. These studies suggest that MRA can be used as an institution's primary diagnostic imaging modality for patients requiring bypass operations, but we must caution that each institution must first develop and document their expertise at accurately interpreting MRA. The decision to operate on the basis of MRA alone must be individualized to each patient, and confirmatory DSA may be indicated, if its findings would allow for a procedure with less risk than distal bypass.

Preoperative arterial imaging techniques must demonstrate arterial stenoses as accurately as patent or occluded segments to offer patients the best option for extremity revascularization. MRA over- and un- derestimated stenoses seen on DSA in equal fre- quency in this study. McCauley et al.10 reported a sensitivity of 67% and a specificity of 89% for MRA detection ofstenoses greater than 75% in a series of 22


D

Fig. 2. Corresponding MRA (A, B, and E) and DSA (C and D) images of patient with diabetes and gangrenous toe who underwent left popliteal-posterior tibial bypass and superficial femoral patch angioplasty. MRA (A, arrow) identified focal superficial femoral artery origin stenosis seen on DSA (C, arrow). Patent posterior tibial artery seen on MRA (B, small arrows) was misidentified as peroneal on DSA (D, small arrow). Diffusely diseased tibioperoneal trunk was identified by both studies (B and D, large arrows). MRA axial images (E) confirmed tibial artery as posterior tibial (arrow), medial to tibia (T) and fibula 09.

JOURNAL OF VASCULAR SURGERY Volume 23, Number 5 Hoch etal. 799

Fig. 3. Corresponding MRA (A, B, C, D) and DSA (E, F, G, H) images of patient with failing femoropoliteal graft. MRA failed to predict appropriate management plan, because critical stenosis near proximal anastomosis seen on DSA (E, straight arrow) was interpreted on MRA as signal dropout caused by surgical clips (A, straight arrow). Distal graft stenosis was correctly identified by both MRA (C, curved arrow) and DSA (G, curved arrow).


patients with peripheral vascular disease. Her tz et al., 12 however, found a high degree of correlation between studies when the severity of stenosis was calculated with a calibrated magnifying glass from the axial MRA images. Primarily projectional images were used in this study; therefore additional review of the axial images for stenotic segments might have improved our accuracy for detecting stenosis.

Despite MRA's modest sensitivity for diagnosing stenoses, the mismatches between MRA and DSA did not adversely affect our ability to plan the appropriate clinical management of extremities with stenotic arteries before operation. This success was partly due to the low incidence of discrepancy between studies in the infrainguinal arteries most suitable for PTA such as the superficial femoral and popliteal arteries. MRA offers patients with threatened limbs or severe claudication who may not be candidates for bypass because ofcomorbid conditions an accurate assessment of their arterial anatomy. In selected patients MRA may allow limb salvage by PTA via an antegrade approach. Antegrade catheterization limits the amount of contrast material and the morbidity. Pa- tients without reconstructible disease on MRA can avoid the morbidity of conventional arteriography. For patients who are not surgical candidates and have severe lifestyle-limiting claudication, MRA offers a risk-free noninvasive technique to determine whether their arterial disease might be treated by PTA.

We did not include our analysis of the pelvic arteries in this report, because with the commercial software available during the time of the study, we were not confident interpreting 2-D TOF MRA of the iliac arteries. The anatomic course of the iliac arteries predisposes them to "in-plane" flow artifacts, result- ing in overestimation of disease. Currently we are prospectively evaluating a combination of MRA techniques to stud), the pelvic vasculature including the use ofnonnephrotoxic MRA contrast with 2-D TOF and phase contrast MRA. ~3,14

At our institution MRA allows an average cost savings of $756 per patient over preoperative imaging with DSA. Carpenter et al. 6 similarly reported a S800 cost savings with MRA compared with outpatient conventional arteriography. Cambria et al. 8 reported a $2000 per patient cost savings with MRA.

In conclusion, we analyzed 50 extremities that underwent both 2-D TOF MRA and digital subtraction angiography. Separate clinical management plans for the ischemic extremities were formulated on the basis of each imaging modality and the available clinical history and noninvasive laboratory findings. MRA- and DSA-based revascularization plans were in agreement in 90% ofischemic extremities. On the basis of these retrospective data, we have embarked on a

prospective protocol, performing preoperative outpatient MRA before rcvascularization and obtaining conventional angiograms at the time of surgery or attempted PTA. We agree with others 6-9 that MRAis a safe, noninvasive, and cost-effective outpatient imaging modality. When used in combination with the patient's physical examination and segmental limb pressures with plethysmography, MRA is sufficient for planning infrainguinal arterial bypass procedures and selecting patients for pcrcutaneous transluminal angioplasty.

REFERENCES

1. Turnipseed WD, Detmer DE, Berkoff HA, Acher CW, Crummy AB, Belzer FO. Intraarterial digital angiography: a new diagnostic method for determining limb salvage candidates. Surgery 1982;92:322-7.

2. Diagnostic imaging in peripheral vascular disease. Kozak BE in basic data underlying clinical decision making in vascular surgery. In: Porter JM, Taylor LM Jr, editors. St. Louis: Quality Medical Publishing, 1994:16-24.

3. O'Moore PV, Denham IS, Steinberg FL, et al. The complications ofangiography: a prospective study [abstract]. Radiology 1988;169:317.

4. Waugh IR, Sacharias N. Arteriographic complications in the DSA era. Radiology 1992;182:243-6.

5. Turnipseed WD, Sproat IA. A preliminary experience with use of magnetic resonance angiography in assessment of failing lower extremity bypass grafts. Surgery 1992;112:664-9.

6. Carpenter JP, Baum RA, Holland GA, Barker CF. Peripheral vascular surgery with magnetic resonance angiography as the sole preoperative imaging modality. J Vasc Surg 1994;20: 861-9.

7. Carpenter JP, Owen RS, Baum RA, et al. Magnetic resonance angiography of peripheral runoffvessels. I Vase Surg 1992; 16: 807-15.

8. Cambria RP, Yucel EI(, Brewster DC, et al. The potential for lower extremity revascularization without contrast angiography: experience with magnetic resonance angiography.

Vase Surg 1993;17:1050-7. 9. Owen RS, Carpenter JP, Baum RA, Perloff LJ, Cope C.

Magnetic resonance imaging ofangiographically occult runoff vessels in peripheral arterial occlusive disease. N Engi J Med 1992;326:1577-81.

10. McCauley TR, Monib A, Dickey KW, et al. Peripheral vascular occlusive disease: accuracy and reliability of time-of-flight MR angiography. Radiology 1994;192:351-7.

i1. Wesbey GE, Higgins CB, Amparo EG, Hale JD, I(aufman L, Pogany AC. Peripheral vascular disease: correlation of magnetic resonance imaging and angiography. Radiology 1985; 156:733-9.

12. Hertz SM, Baum RA, Owen RS, Holland GA, Logan DR, Carpenter JR Comparison of magnetic resonance angiography and contrast arteriography in peripheral artery stenosis. Am J Surg 1993;166:112-6.

13. Prince MR, Yucel EK, Kaufman JA, Harrison DC, Geller SC. Dynamic gadolinium-enhanced three-dimensional abdominal MR arteriography. Magn Reson Imaging i993;3:877-881.

14. Swan JS, Grist TM, Weber DM, Sproat IA, Wojtowycz MM. MR angiography of the pelvis with variable velocity encoding and a phased-array coil. Radiology i994;190:363-9.

Submitted Oct. 11, 1995; accepted Feb. 4, 1996.

JOURNAL OF VASCULAR SURGERY Volume 23, Number 5 Hoch et al. 801

D I S C U S S I O N

Dr. John L. Glover (Royal Oak, Mich.). Dr. Fioch and his associates have shown us that magnetic resonance angiography can be used in planning the treatment for infrainguinal disease in most patients, therefore avoiding the risks of conventional angiography for diagnostic pur- poses.

I have four questions and one comment, Dr. Hoch. First of all, did you use any of the MR contrast agents? If so, tell us about that. Second, was your series a consecutive one, or were the patients selected in some fashion? If so, please let us know. Third, all of your patients had normal or near normal inflow. Do you have any information about the effect of aortic iliac disease on the accuracy of magnetic resonance angiography? The final question concerns a patient, a 70-year-old with long-standing insulin-dependent diabetes and tissue loss with exposed bone in the forefoot, but no rest pain. He has normal flow to both popliteal arteries, but all of the calf vessels occlude within a few centimeters of their origin. Neither MRA nor conventional angiography show flow at the ankle or the foot. How would you manage such a patient? That case leads to my comment. In both your study and in a prospective multicenter comparison of MR angiography and conventional angiography, there is a percentage of patients where there is no accurate information regarding flow at the ankle or the foot. There is a little flow with both studies but no information about the character of the vessel wall. On the other hand, duplex scan can give information about both those features, and I think it can be used in some cases to plan treatment without conventional or magnetic resonance angiography. I would appreciate your comments about your use of duplex scan in these kinds of patients.

Dr. John R. Hoch. Relative to your first question, whether we used nonnephrotoxic contrast agents in this study, we did not. This study was designed retrospectively to evaluate patients who primarily present with infrainguinal occlusive disease. We currently have underway, in our institution, several prospective studies evaluating the problem of MRA imaging of inflow arteries. As most of you know, MRA has a problem with diagnosing stenoses in the iliac arteries because of" in-plane" flow artifacts. Intraarte- rial use of the agent gadolinium or one of its derivatives offers an advantage in that it shortens the T1 relaxation of blood, causing a very high signal in the blood compared with the background tissues. Gadolinium, as a "contrast" agent, has the potential to eliminate the flow phenomena limitations of standard 2D time of flight MRA techniques. Presently, we are in a clinical protocol using gadolinium with a Tl-weighted 3D gradient echo technique. I f my slides were still available, I could show you some very interesting MRA images using this protocol, which demonstrate elimination of "in-plane" flow artifacts and improved visualization of the renal arteries using gadolinium.

The second question was relative to our patient selec- tion. Again, this was a retrospective review and suffers from the problems of all retrospective reviews. Patients were selected for inclusion who had palpable femoral pulses and the signs and symptoms ofinfrainguinal peripheral vascular disease, confirmed by PVR findings.

Regarding your patient, in whom imaging revealed an open popliteal system with loss of tibial vasculature at the level of the ankle without vessel visualization in the foot, we do not have experience with Duplex imaging of the pedal arteries. Two-dimensional time-of-flight MRA has the capability to detect flow as low as 2 ml a minute, far better than conventional contrast angiography. I would bring this patient back to the MRA suite and take a more focused look at the foot with the extremity coil, hoping to identify a vessel suitable for bypass. This sort of patient highlights the necessity for good communication between vascular surgeons and our radiology colleagues to plan imaging studies specific for patients' disease processes.

Dr, John Blebea (Cincinnati, Ohio). I believe that the place of MR angiography is now much better defined in infrainguinal disease. With the recent publication last week of the results from the multicenter prospective trial on 155 patients (JAMA 1995;274:875-80), it has been demonstrated that MICA and standard contrast angiography are equivalent in diagnostic accuracy in the leg. Aorto-iliac inflow disease, on the other hand, can be problematic, and its diagnostic usefulness is more dependent on the expertise available at each institution. Our radiologists are now investigating the efficacy of both gadolinium contrast infu- sion and cardiac gating to improve accuracy in this area. If such technologic advancements can improve imaging to what we now have in the leg, I believe MRA will find an important place in the routine evaluation of such patients.

Dr. Hoch. Dr. Blebea was one of the principal authors of a paper entitled, "Multicenter Trial to Evaluate Vascular and Magnetic Resonance kalgiography of the Lower Ex- tremity," which was recently published in the September 20th issue of JAMA. This is a very interesting study, in that it prospectively compared magnetic resonance angiography of the lower extremity wit h conventional angiography. This was a multicenter study in which the University of Wiscon- sin participated. This study found that the sensitivity for distinguishing between patient and occluded aa'terial segments was 83% for conventional angiography and 85% for magnetic resonance angiography. Both studies had an 81% specificity. Comparing the proposed patient treatment plans in this study, there was an 87% agreement between MRA and conventional angiogram plans, whereas in 13% of the cases, the treatment plan differed. It was the conclusion of the authors that the addition of magnetic resonance angiography to treatment plans derived from the conventional angiograms clearly improved patient outcomes.

Documents

Use of magnetic resonance angiography for the preoperative evaluation of patients with infrainguinal arterial occlusive disease